INDEX

Explanations

phrases indicating widespread international or national coverage

oai_token-act-pair · gpt-3.5-turbo

references to global locations and notable places

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Cosponsors

-0.78

ierre

-0.77

xit

-0.75

nect

-0.74

aroo

-0.72

uci

-0.72

omas

-0.69

icho

-0.68

vg

-0.68

ulhu

-0.67

POSITIVE LOGITS

 rainbow

0.86

 aisle

0.82

 nodd

0.79

 corners

0.73

 fuss

0.72

igree

0.71

 dozen

0.71

 arena

0.71

 imaginable

0.68

 globe

0.67

Activations Density 0.111%

phrases indicating widespread international or national coverage

references to global locations and notable places

No Comments

No Known Activations